Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
202 Open 1,502 Closed
202 Open 1,502 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Refactor: Decouple Core Transformer Blocks
#1852 opened Jun 19, 2025 by parambole Loading…
4 tasks done
@RissyRan @gobbleturk
Fork Splash Attention kernel to MaxText
#1851 opened Jun 19, 2025 by copybara-service bot Loading…
1
Allow moe_test.py to be run on internal tools.
#1847 opened Jun 18, 2025 by copybara-service bot Loading…
1
Resolve pylint errors in moe.py.
#1846 opened Jun 18, 2025 by copybara-service bot Loading…
1
Refactor input pipeline with synthetic data iterator
#1845 opened Jun 18, 2025 by NuojCheng • Draft
4 tasks done
Get Linter / CPU tests to succeed
#1844 opened Jun 18, 2025 by SamuelMarks Loading…
4 tasks done
15
Add paged attention into DeepSeek v3
#1843 opened Jun 17, 2025 by copybara-service bot Loading…
1
Enable Checkpoint Conversion from Huggingface to Maxtext
#1839 opened Jun 16, 2025 by YixuanWang-99 • Draft
4 tasks
Integrate Multi-Token Prediction (MTP) Training objective
#1837 opened Jun 16, 2025 by parambole Loading…
4 tasks done
@richjames0 @mailvijayasingh
@RissyRan @hengtaoguo @gagika @gobbleturk @A9isha
pyconfig → pydantic
#1836 opened Jun 15, 2025 by SamuelMarks • Draft
4 tasks done
Add Qwen3
#1835 opened Jun 14, 2025 by bzantium Loading…
4 tasks
1
Refactor profiler in trainers pull ready
#1833 opened Jun 14, 2025 by SurbhiJainUSC Loading…
4 tasks done
9
JetStream Offline Engine
#1829 opened Jun 13, 2025 by wenxindongwork Loading…
4 tasks done
Upgrade dependencies (now supports and defaults to Python 3.12 and Ubuntu 24.04)
#1826 opened Jun 12, 2025 by SamuelMarks Loading…
4 tasks done
3
Moving device_put for the kernel earlier to avoid a host convert op
#1824 opened Jun 12, 2025 by zhenying-liu Loading…
Use the extended jax.experimental.colocated_python.colocated_cpu_devices API
#1822 opened Jun 11, 2025 by copybara-service bot Loading…
fix runtime errors in colocated python dataloader
#1819 opened Jun 11, 2025 by sadikneipp Loading…
4 tasks done
1
[WIP] llama4 ckpt conversion
#1816 opened Jun 10, 2025 by aireenmei • Draft
4 tasks
Refactor: Recording and logging training and evaluation metrics in all trainers
#1815 opened Jun 10, 2025 by SurbhiJainUSC Loading…
4 tasks done
@richjames0 @A9isha
12
Add total memory usage logging in GB and unit tests for activation offload pull ready
#1813 opened Jun 10, 2025 by zhenying-liu Loading…
9
[WIP] Full Llama4 VisionModel
#1809 opened Jun 5, 2025 by hengtaoguo • Draft
4 tasks
fix github clone link
#1807 opened Jun 5, 2025 by junjieqian Loading…
4 tasks done
4
I've added unit tests for critical conversion, layer, and inference c…
#1804 opened Jun 4, 2025 by stingram • Draft
1 of 4 tasks
[NVIDIA] Add packing support for GPU flash attention
#1802 opened Jun 4, 2025 by kocchop Loading…
4 tasks done
[DRAFT] v7x AOT support
#1796 opened May 31, 2025 by shralex Loading…
4 tasks done
ProTip! Type g i on any issue or pull request to go back to the issue listing page.